Inching Towards A Targeted Therapy for Preeclampsia

Novel drug development for pregnancy specific conditions remains a challenge. The reasons for this sentiment are many, but lack of a sound understanding of the pathogenesis for various conditions in pregnancy ranks high among them. Preeclampsia, one of the most common medical complications of pregnancy, has thus far eluded comprehensive biological understanding.¹ In the absence of a sound understanding of its biological underpinnings, our first attempts to alleviate the consequences of this disorder targeted the classic signs and symptoms of preeclampsia (hypertension and proteinuria) with agents such as anti-hypertensive compounds.¹ Trials of these agents – numbering over a dozen – failed to demonstrate that control of blood pressure itself could reduce the systemic consequences of preeclampsia. Nevertheless, lack of success in these early efforts did not deter further attempts, and primary prevention trials premised on biochemical differences in women with and without preeclampsia represented the next generation of such clinical trials.

Low urine calcium and low calcium intake noted in preeclampsia, in combination with the vasodilatory activity of calcium itself, prompted primary prevention trials with calcium supplementation.¹ Similarly, imbalances between prostacyclin and thromboxane led to large trials examining the utility of aspirin to prevent preeclampsia.¹ Primary prevention trials have been informative, and given that the incidence of preeclampsia is approximately 4–7% in nulliparous women, these have necessitated inclusion of thousands of women enrolled over an extended period of time. While noteworthy in their magnitude and outstanding execution, these trials have, for the most part, been neutral or negative. Primary prevention studies by their nature require large numbers of subjects, especially in the setting of outcomes that are relatively infrequent. Furthermore, given the history of these trials in preeclampsia, mustering the appetite to initiate yet another trial of similar nature – unless upcoming primary prevention trials reveal surprises (e.g., clinicaltrials.gov NCT00135707) – may represent a potentially insurmountable challenge.

Without a sound understanding of the underlying pathophysiology, therapies focused on pathways thought to be involved in the pathogenesis of preeclampsia will remain non-specific. The pathogenesis of preeclampsia is a two-stage process – the first is an asymptomatic stage that involves abnormal placentation (placental stage), and the second is placental elaboration of soluble factors that enter the maternal circulation and lead to widespread endothelial dysfunction (maternal stage). Soluble Fms-like tyrosine kinase-1 (sFlt-1) is a factor thought to be involved in the second state of this process. sFlt-1 is a natural inhibitor of vascular endothelial growth factor (VEGF) and placental growth factor (PlGF), both of which are present in maternal circulation during normal pregnancy. Recent data indicate that sFlt-1 levels are markedly elevated in patients diagnosed with preeclampsia,² and sFlt-1 over expression leads to hypertension, proteinuria, and glomerular endotheliosis, the hallmarks of human preeclampsia.³ This concept was advanced in proof of concept intervention studies by Li, et al.; VEGF-121 administration in Sprague-Dawley rats with elevated sFlt-1 levels successfully ameliorated all three features of the disease,⁴ suggesting this could represent a potential therapy in women with severe preeclampsia.

Proof-of-concept studies have not halted. In fact, the manuscript by Gilbert et al. published in this issue of Hypertension advances the existing anti-angiogenic hypothesis.⁵ These investigators pioneered a novel model of preeclampsia in rats that involves induction of utero-placental ischemia in the second trimester. Utero-placental ischemia may be the critical initiating step in the pathogenesis of preeclampsia. These investigators previously demonstrated that this model develops several of the features of preeclampsia including hypertension, fetal growth restriction, and decreased glomerular filtration rate.⁶ Of note, however, is the lack of significant proteinuria and glomerular endothelial histological changes in this model – characteristic findings in human preeclampsia. This absence may be due to the limited exposure (2–3 days) of toxic mediators to maternal vasculature. This contrasts with the model involving induced uterine artery ischemia in non-human primates where, in addition to gestational hypertension, there is definite proteinuria and glomerular endotheliosis.⁷ Nevertheless, Gilbert, et al. previously reported that their model is associated with increased circulating levels of sFlt-1 and, building on prior work by Li, et al. (above), they convincingly demonstrated improvement of hypertension and glomerular filtration rate by VEGF-121. The VEGF-treated animals also improved the abnormal vascular reactivity noted in this model. Importantly, this therapeutic strategy appeared not to impose any adverse outcomes on the fetuses, in fact, fetal weights may have improved. Thus, in the setting of several pathways potentially altered in this model of preeclampsia, one intervention appeared to have a biologically and physiologically important effect. Based on this study and that from Li, et al., are we prepared to move forward to human trials with this single target?

With nearly 25,000 entries about preeclampsia in PubMed, the debate about pathogenesis and potential biomarkers remains in full force. Are we ready to utilize the markers thus far identified to inform clinical trials? Our understanding of the pathogenesis of preeclampsia has undergone tectonic advances in the past decade.⁸ Uncovering potential pathogenic mechanisms may satisfy some, but if we never build upon these advances to develop therapies, we will merely continue to debate hypotheses without ever making a difference to our patients – both mother and baby. The first criticism in moving any hypothesis forward in preeclampsia will be whether the target of choice is the right target, especially given the number of alterations noted in women with preeclampsia.^{1, 8} Indeed innumerable alterations in preeclampsia have been described. Lessons from inflammatory arthritis, however, may help guide the preeclampsia community in such a situation.

When Feldmann and Maini decided to attack the proinflammatory cytokine storm in rheumatoid arthritis with a single agent, the prevailing view was that cytokines would not make good therapeutic targets. Targeting a single cytokine was believed to be ineffective, while blocking many cytokines was deemed impractical.⁹ Feldmann and Maini targeted one specific cytokine, TNF-α, because experimental studies suggested TNF-α inhibition would inhibit several downstream cytokines thought to be intimately involved in the inflammation and tissue destruction characteristic of rheumatoid arthritis. The cytokine considered most indicative of inflammatory arthritis was IL-1. When screening potential agents, their primary read-out was inhibition of IL-1 production. While other pathways almost certainly contribute to rheumatoid arthritis, the inhibition of a single pathway was sufficient to significantly alter the signs and symptoms of this devastating condition. Similarly, a multitude of pathways in preeclampsia may exist, but can we use sFlt-1 to guide the development of therapeutic agents in preeclampsia?

While the debate about the pathogenesis of preeclampsia marches on, and frankly may never end,¹⁰ biomarkers – whether they are mediators or simply biomarkers of the disorder – can be used to aid in our quest for therapies. Short term proof-of-concept trials with an intermediate outcome (reduced serum levels of sFlt-1 or a rise in PlGF levels) may allow therapies to be tested in a shorter period of time. If a therapy appears not to effect potential intermediates such as sFlt-1, its value should be reconsidered. Prostate specific antigen (PSA) itself is not likely to be pathogenic for prostate cancer, yet this biomarker does inform intervention trials about efficacy, remission, and prognosis. The same can be said of troponin and hsCRP levels for cardiovascular disease. In the case of sFlt-1 not only do epidemiological studies suggest it may be useful biomarker, sFlt-1 (unlike PSA, Troponin, and hsCRP) is likely involved in the pathogenesis of the disease. Therefore, another approach may be to develop therapies specifically targeting sFlt-1.

Therapeutic options include administration of small molecules that block sFlt-1 production, recombinant ligands for sFlt-1 (e.g., VEGF as in the case of Gilbert et al.), or neutralizing antibodies against sFlt-1 and/or angiogenic proteins. Alternatively, we could identify means of removing sFlt-1. sFlt-1 levels during normal pregnancy can reach levels ~50 fold higher than levels seen in non-pregnant women,³ and levels increase even further among patients with preeclampsia.² The function of sFlt-1 in “normal” pregnancy is not entirely known. sFlt1 inhibits trophoblast migration and differentiation in vitro, and therefore it has been hypothesized this “normal” production of an anti-angiogenic protein is nature’s way of slowing and reversing the placental angiogenesis that is a hallmark of normal mammalian pregnancy. Nevertheless, the goal for any intervention should be to bring levels back to a “normal range” in the event that a lower limit may be necessary for normal pregnancy and delivery to proceed. The study by Gilbert et al. provides further support for targeting sFlt-1 in women with severe preeclampsia. The natural next steps of course would be to choose a safe and effective strategy and test this in non-human primates. If successful, the therapy could be moved to women with severe preeclampsia, particularly pre-term preeclampsia. While we continue to debate which pathway is most relevant, and thus which specific one to target, targeting one pathway with strong biological support can have a marked effect on an otherwise complex disorder.